ΣΥΜΦΩΝΟΙ ἈΡΙΘΜΟΙ: A Note on Republic 531C1-4

Author(s): Andrew Barker

Source: Classical Philology, Vol. 73, No. 4 (Oct., 1978), pp. 337-342
Published by: The University of Chicago Press
Stable URL: https://www.jstor.org/stable/268854
Accessed: 15-03-2020 05:15 UTC

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

The University of Chicago Press is collaborating with JSTOR to digitize, preserve and
extend access to Classical Philology

This content downloaded from 99.238.17.184 on Sun, 15 Mar 2020 05:15:30 UTC
All use subject to https://about.jstor.org/terms
 NOTES AND DISCUSSIONS

2TM4QNOI 'APIOMOI: A NOTE ON REPUBLIC 531C1-4

rois -yp Ev TabTaLs TaZs a/vpLOvaLas Tcas &KOvoivUats &pLpOJobs ?7oOatv, \XX'
rpo,3Xicuatra LvtaLwLV, i7rtOKO7rELv TlVES 0TiU4LcJoV0 &pLO/,UO0 KaC TIVJES ov, Kat 6L( 7I E

I do not intend to make any significant comment on the first part of the passage,
though its plain sense will be involved in my remarks about its latter portion,
beginning at r-TaKo7rEZv rTves. Plato's complaint, in broad terms, is that the Py-
thagoreans of his day, though they are indeed looking for apLouoi, seek them in the
concords which they hear, rather than using the evidence of their ears to establish
abstract problems for the mind to resolve.' My interest here lies in his specific
recommendation that they should ask Tives Lc/XOVoL apLO6,U0t Kac TIVJES 0o, Kat &a&
EKaTrepo.

What sort of thing had Plato in mind? To begin from the end, he should clearly
not be requiring at this point in his argument that they demonstrate the "con-
cordance" of certain numbers from metaphysical first-principles. That can only
come later, as a product of dialectic: OKOUV oVO ros '577 abr6s iaTLV 6 6,uos Ov rb btaX\iyEOc
lrepaLivEL; The studies which form the rpootuov to dialectic seem not to search more
deeply into reasons and causes than the procedures of mathematics can by them-
selves allow.2 Yet it is hard to see what might count as a mathematical reason why
certain numbers are T/WOvOVOL, at least if part of what Plato means is that actual
musical sounds corresponding to or participating in these numerical "forms" will
themselves be concordant.
Second, and more centrally, what are the numbers for which Plato is asking
the Pythagoreans to search? An obvious suggestion would be that they are the
ratios by which it is possible to represent the various kinds of musical interval.
But whether these are to be thought of as ratios (of lengths of strings, of speeds
of propagation, of frequencies of vibration), or as purely abstract,3 the general
principles had long since been adopted by the Pythagoreans, and the basic ratios
identified. All this was and is well known. There is nothing new in Plato's day in
the association of the ratios 2:1, 3:2, 4:3, and 9:8 with their respective intervals:

1. Rep. 531B-C with 530A-B and 523A-524D.

2. Rep. 531C9 ff., especially 531E4-532A2.
3. The association of the ratios with lengths of strings appears to have been early Pythagorean
practice, though it is hard to find an authoritative statement to this effect in the sources. Other
physical magnitudes were also related in this way: see D.-K. 18, passages 12 and 13. Cf. also the
use of the word KacVbv in the title of the Euclidean Kararoj.i KapoVos (Sectio canonis). Both Plato
(Tim. 80A-B) and Aristotle (e.g., Gen. an. 786b7 ff.), perhaps following Archytas (D.-K. 47 Bi,
A19a), at times linked the ratios with the speed of a sound's propagation: their connection with
frequencies of vibration seems to be due to Heraclides Ponticus (see Porphyrios' Kommentar zur
Harmonielehre des Ptolemaios, ed. I. DUring [Goteborg, 1932]<, 29. 27-30. 21, and DUring's "Ptole-
maios und Porphyrios iiber die Musik," Gdteborgs Hogskolas Arsskrift 40 [19341: 161 ff.). Plato com-
monly has in mind a more abstract view, as the nature of our present passage would suggest. Cf.
Phil. 16-17. On the question generally, and Heraclides in particular, see K. von Jan, Musici Scrip-
tores Graeci, (Leipzig, 1895), 1:134-41: cf. also the Theophrastus fragment in Porphyry's Com-
mentary, 61. 22-65. 15 (During) = frag. 89 (Wimmer).

Permission to reprint a note in this section may be obtained only from the author.

337

This content downloaded from 99.238.17.184 on Sun, 15 Mar 2020 05:15:30 UTC
All use subject to https://about.jstor.org/terms
 338 NOTES AND DISCUSSIONS

it is explicitly at work in the Timaeus (35-36 and 43D), and, as Adam's note indi-
cates,4 it may well lie behind the sense of the vexed passage at Repu5lic 400A5-6.
Besides, what can Plato have intended other than these number-ratios and their
like when he asserts that the Pythagoreans iv raTrats rats av,owvIats rats aKOvouiVatf
AptO,uobs rr7roDatlo?
The possibility which springs most readily to mind is that what Plato is after
is not just the relevant ratios, but an indication and an understanding of the
mathematical classes to which the ratios belong. As they stand they are simply
pairs of numbers: what distinguishes them, for instance, from the ratios 5:3 or
9: 7? In the introductory passage of the Euclidean Sectio canonis the author makes
just such a classification. All concords fall under ratios which are either 7roXXcaLrXaLos
or b-ru6ptos, never brtgpis. That is, in the pair of numbers forming a ratio corre-
sponding to a concord, one is either a multiple of the other, or it stands to it in
superparticular ratio: x = (1 + 1/n)y. What it cannot stand to it in is super-
partient ratio, i.e., x = (1 + p/n)y, where p is greater than 1 and smaller than n.5
The attribution of the Sectio to Euclid has been questioned,' but not, I think,
the relatively early date of the major part of it. It seems to be just the sort of
thing which in his own time was making Aristoxenus spit fire.7 Unfortunately, the
same cannot be said for the short preliminary section, from which the classification
I have mentioned is taken: K. von Jan, for one, assigns it-rightly, in my opinion-
to a lesser author and a later date.8 Nevertheless, its author was intelligent enough
to realize that the distinctions he marks are ones which the main arguments of
the work cannot do without, and there can be no doubt that whoever wrote the
Euclidean passages was familiar both with the classification and with the assign-
ment of concords to two of the classes only-so familiar, perhaps, that he con-
sidered it not worth his while to argue the point explicitly.9
It seems plausible to suppose that the classification of the Sectio is just the sort
of thing for which Plato was looking. Of course it must on any sensible dating be a
good deal later than the Republic, so that Plato might still have been right in
treating the subject as terra incognita in his day. But three pieces of evidence
seem to count against the hypothesis.
(1) Nicomachus, in his Introductio arithmetica, distinguishes five kinds of ratio,
among which our three are included.'0 (The other two import no new ideas.)
Nicomachus is a late source and a fairly incompetent mathematician, but I am in
no position to dispute Heath's judgment that for Nicomachus' work, as for Theon
of Smyrna's and lamblichus', "there is little in their content that does not go back
to the immediate successors of Pythagoras."" If the classification of ratios goes
back so far, Plato could hardly be ignorant of it, or of its musical application.

4. James Adam (ed.), Plato: "The Republic"2 (Cambridge, 1963), ad loc. I am not altogether
happy with Adam's conclusion. Can the ratio 9:8 be thought of as being as fundamental as the
others, on the same systematic plane? It appears to be derivable. Cf. D.-K. 47 A16 (Archytas),
Eucl. Sect. can. 13, and even, in his own very different way, Aristoxenus El. harm. 21, 46, 55-58.
5. Eucl. Sect. can., introduction (pp. 23-24 Meibom, p. 149 von Jan).
6. For various suggestions and arguments see T. L. Heath, The Thirteen Books of Euclid's "Ele-
ments" (Cambridge, 1908), 1:17.
7. E.g., Aristoxenus El. harm. 32.
8. Von Jan, Musici Scriptores Graeci, 1:117-19.
9. The relevant assumptions are pervasive, but see especially 10-12.
10. Nicomachus Intr. arith. 17 ff.
11. T. L. Heath, Greek Mathematics (New York, 1963), p. 61.

This content downloaded from 99.238.17.184 on Sun, 15 Mar 2020 05:15:30 UTC
All use subject to https://about.jstor.org/terms
 NOTES AND DiscussioNs 339

(2) Archytas, with whose work Plato was certainly familiar,12 is known to have
worked out the numerical representations for the intervals of the tetrachord in
each of the three yi7,.13 This places his activity in the general area of mathematical
harmonics. There has also been preserved his proof of the proposition that there
can be no number "which is a [geometric] mean between two numbers in the ratio
known as e7-ru6ptos or superparticularis, that is, the ratio (n + 1): n."'4 As Heath
points out, the proposition (and in most respects its proof) is the same as that
set out in the Sectio, where it forms a crucial part of the argument through
which the divisions of the scale are derived.15 We must, I think, judge it highly
probable that the use to which Archytas put it was the same, and that the dis-
tinctions between the kinds of ratio and the ways in which these distinctions are
to be deployed in musical theory were in fact being mapped out, not merely
contemporaneously with Plato, but by the self-same Pythagoreans whom he is
apparently attacking, and under his very nose."6 All of which adds force to the
remaining and somewhat vague consideration to which I want to refer.
(3) Immediately after the remarks of Socrates which form the subject of this
note, we get Glaucon's response: Aa/g6vLtov -y6p, ef7 wpary,.ca Xi-yets, a remark which I
take to indicate both the excellence and elevation, and the extreme intellectual
difficulty, of such studies. (Compare the notion of the bxtcuo6os &ivrp in a similar,
though more literally "daemonic" context, at Symp. 203A5.) Socrates does not
dissent, indeed by implication agrees, and goes on to commend these studies'
utility: and it would, after all, be most surprising if he were to dispute the con-
tention that they require quite unusual intellectual talent and application, since
a discipline which involved no more than the grasp of a few simple distinctions
and some relatively unproblematic theorems concerning them would scarcely
have suited his curriculum."7 If even the general principles involved in the research
he is commending were well understood in his own day and by his own friends,
his remarks would have appeared strangely inept. He would have been commending
as novel, remarkable, and un-Pythagorean a branch of research which he already
knew to be a part of the Pythagorean achievement. This is unbelievable. Whatever
Plato was after, it cannot have been what the Pythagoreans were already doing.
What then can it be which Pythagoreans were not doing, and which Plato saw
as essential? We may reasonably suppose that the study of classes of ratio was in
Plato's mind; and, if Archytas' derivation of the intervals of the tetrachord
depended directly, like that of the Sectio, on theorems concerning these classes,
12. D.-K. 47 Al. A2, A5.
13. D.-K. 47 A16.
14. D.-K. 47 A19. See Heath, Mathematics, p. 136.
15. Eucl. Sect. can. 3 and (e.g.) 16.
16. The relation between Plato and Archytas deserves careful study, which I cannot give it her
As far as the present point is concerned, unless we suppose their friendship and Plato's knowled
of the Pythagoreans of southern Italy not to have begun until after the Republic was written,
presumably during Plato's second visit to the West, my contentions seem to hold. The close rese
blance often noted between Plato's orOXLTEIa and the Pythagorean communities, apart from an
external evidence, makes this very unlikely; and several authors have remarked on the accurac
with which Plato's mathematical curriculum reflects Archytas' list of sciences to which number
provided the key (D.-K. 47 Bi).
17. It is of the essence of the curriculum that it should exercise the mind in hard work,
521-524. But of course, if this difficulty could be avoided, the evidence which I have moun
cast doubt on the hypothesis that it is the classification of ratios which Plato has his eye on a
becomes evidence in favor of that hypothesis. Did Plato consider Archytas' research, eit
itself or in its potential for further study, to be anything so remarkable as to be called 3aqu

This content downloaded from 99.238.17.184 on Sun, 15 Mar 2020 05:15:30 UTC
All use subject to https://about.jstor.org/terms
340 NOTES AND DISCUSSIONS

it must have seemed to Plato a step in the right direction. The particular intervals
involved in the musical scale could now be determined by deduction from mathe-
matical premises, rather than by empirical experiment alone.'8 But Plato apparent-
ly wanted more. Was it that he wanted the concords and discords pinned down
still more closely, and assigned accurately to classes of ratio commensurate with
themselves (for not every brtu6pLos ratio corresponds to a concord)? If so, tidy
though such a scheme would be, it is not clear why it is needed, given that Archytas
and subsequently Euclid were able to derive the intervals of the scale using no
more than the existing classifications. Could it be that what he required was
nothing less than an account of the reasons why these classes of ratio are associated
with these classes of interval? Yet these reasons, as we saw at the start, must
inevitably take us beyond the scope of mathematics and into the province of
dialectic.
The purpose of this note was to raise the difficulty. I am not at all sure that I
can solve it: but a suggestion may be in order. It is evident that Plato sees his
objection as somehow dependent on the excessive empiricism which he attributes
to the Pythagoreans. That is the nub of the comparison with the astronomers at
531B8-C1, and it determines the way in which the whole topic is introduced,
including the digression concerning those who go much further along the wrong
road than the Pythagoreans, using their ears in place of their minds. Plato gives
space to these "extremists" presumably in order to put into our heads the kind
of criticism he is concerned to make of the Pythagoreans, as well as to mark a
contrast between the two schools: and perhaps the comparison with astronomy
indicates the real point. The Pythagoreans undoubtedly approached that subject,
too, with number in mind, but seem not to have come so close to a complete
mathematization of it as they did with music. The case of astronomy is therefore
relatively easy: those whom Plato attacks are interested in discovering and tabulat-
ing the numerical relations between movements of visible objects,19 but not in
attempting to reduce the study of them to an investigation of the properties of
numbers alone. In music, as we have seen, some such attempt does appear to
have been made. Nevertheless, Plato may be suggesting, the basis on which it has
been made is wrong: it will be found on inspection to be as empirically based as
astronomy, though not so obviously. Hence it is helpful to indicate the parallel,
since the plainer case of astronomy may show up the hidden defects of Pythagorean
musical analysis.
Now this suggests that, if Plato was indeed aware of an attempt by Archytas,
on the lines of the later Sectio, to derive the sizes of the concords and the other

18. Let us grant that it is not possible validly to proceed from purely and exclusively mathematical
premises by deduction to conclusions about concords and discords. Nevertheless, it is such a project
as this that the author of the Sectio believes himself to be engaged in, and in which-by the supres-
sion of important linking assumptions-he gives the impression of succeeding. In the likely event
that the Sectio's arguments are modeled on those of Archytas (see, e.g., E. A. Lippman, Musical
Theory in Ancient Greece [New York and London, 19641, pp. 16-17, and Heath, Mathematics, p. 136),
he may well have presented the same illusory appearance.
19. On Pythagorean astronomy, see, e.g., Arist. Met. 986a6-12 and De caelo 293a20-28. Aristotle,
of course, finds their doctrines insufficiently empirical: but the point is that, however heavily they
depended on appeal to voiyral a'pXaez for the derivation of their propositions, these propositions were
supposed to describe features of the physical universe, even where, as in the case of the supposed
&vrLX0cw, they are actually invisible to us.

This content downloaded from 99.238.17.184 on Sun, 15 Mar 2020 05:15:30 UTC
All use subject to https://about.jstor.org/terms
 NOTES AND DISCUSSIONS 341

intervals of the scale from purely mathematical premises, he was suspicious of

its foundations. Such suspicions would be well placed, since the identification of
the sonorous concord of an octave, for instance, with the ratio 2: 1 cannot follow
from mathematical principles alone. At some point assumptions based on acoustic
experience have to be smuggled in (see n. 18), and that is precisely what Plato
wishes to avoid. But then the same difficulty must pass to him. What exactly is
it that he wishes his mathematical musicologists to do?
If we suppose, as we clearly should, that Plato found the Pythagorean researches
impressive and fruitful, his criticisms cannot seriously touch their procedure, and
must be restricted to their conception of the ways in which their results were to
be explained and justified. It is true that their procedure begins from acoustic
experience: but so too does Plato's.20 Their mathematics itself is not called into
question. But when the ratios corresponding to the divisions of the scale have
been derived and set out, how do we know that they are correct, and what do they
represent? For the Pythagoreans it is the nature of the sounds themselves which
the numbers express:21 and to show that they have got the numbers right they can
only return to experiments with the lengths of strings, together with an Aristo-
xenean appeal to what we find, through perception, to sound concordant or melodic.
The association of heard concords with certain classes of ratio is the basis and the
objective of the whole project, and it is therefore impossible at any stage to abandon
all appeals to the senses.
For Plato, on the other hand, numbers do not directly express the nature of
the sounds we hear. They form a system which the sounds can at best only crudely
reflect. I suggest that the problem which Plato sets his students is to show, without
benefit of sense perception or any appeal to the music which our ears recognize,
how it is that the numbers associated with the notes of the scale form in their own
mathematical right a coherent single system, to the exclusion of other numbers
which would "divide the canon" in other ways. Some such project seems to be
envisaged in the Philebus (16-18), where it is represented, in part, as a search for
an understanding of how music is one, and what the principles of its unity are.
The point is that for Plato, at any rate in the mood of the Republic, such unity
cannot be demonstrated by turning back to the world of the senses, by showing
that the system is a unified whole in that it is what our ears recognize as musical.
Having reached our scheme of ratios, the object is no longer to show that it repre-
sents the series of heard notes in an accepted scale. Its unity and its nature as a
representation of order in the pattern of things cannot arise from the fact that it
represents a single kind of sensed phenomenon, music as heard: we must see it
simply as a system of ratios, and seek ways in which, strictly in its role as a set of
mathematical relations, it constitutes a unity based on an ordering principle which
excludes other forms of numerical proportion. The existence of such a unifying
order as a determining feature of the universe will stand as a ubrMaTLS, enabling us
at the level of &iazvota to derive further theorems and conclusions-about "rational
harmony," not about sonorous music: thus the present topic of research will
20. Rep. 523-524.
21. In addition to our present passage of the Republic, see the general account of Arist.
985b23-986al2, and the alternative interpretations of such number theories as applied t
set out by Theophrastus in the fragment cited in n. 3.

This content downloaded from 99.238.17.184 on Sun, 15 Mar 2020 05:15:30 UTC
All use subject to https://about.jstor.org/terms
 342 NOTES AND DISCUSSIONS

remain, as it should, at the "hypothetical" l

understands them. The status of our principle will not be finally confirmed until
through dialectic we have come to understand the whole nature of unity itself.22
All this is conjectural, and very difficult. If it is near the truth, it suggests that,
so far as they had gone, Plato had no serious quarrel with the Pythagoreans at all.
His complaint is that, having got so far, they go no further: they interpret their
results as a set of truths concerning the sounds we hear; and they seek confirmation
of these results either not at all or merely through a return to their empirical
starting point. They have failed to go on to convert their mathematical representa-
tions of sound into a scheme of purely numerical relations, and to seek the principle
of its unity, what makes it an obala and not a hodgepodge, at the level of mathe-
matics itself. The Pythagorean view can be expressed by saying that mathematics
reveals directly the nature and relations of musical sounds; and that, because the
sounds form an orderly system in the acoustic realm, the numbers which they
essentially are will appear as an orderly system. For Plato the system of numbers
is something merely suggested to the mind by the physical relations of heard sounds
and their means of production: music as heard is patterned, but only very im-
perfectly, on the orderings of numbers, and that order is ultimately a mathematical
form of unity, a numerical offLs, not a musical one.
If this general scheme of interpretation is correct, the expression fT,fb/POL ApLO/.OL'
should not primarily mean "numbers corresponding to the consonances of heard
sound," since it is impossible without relying on the empirical to show that these
numbers and these sounds actually are associated. It is to be construed meta-
phorically, as meaning "those numbers which, from their place in an intelligible
system, we shall call 'consonant' on the analogy with heard sounds." We should
not allow any closer relation between heard musical scales and the numbers of
"musical" science than that a study of the former psychologically suggests the
latter. I am not sure that Plato saw this, at least until the Timaeus-that is, that
he faced straightforwardly the fact that, if we can reach no certain truths about
perceptibles, we therefore cannot be certain that the intelligible truths which we
reach on the prompting of our experience of perceptibles are those same truths
on which the nature of the perceptibles is distantly based. The laws of psychological
association are not so simple. Thus the thesis that the numbers discovered by
these means are the basis for, or are actually represented by, the sounds we hear
must always remain no more than conjectural, and such speculations must be
granted only the rather whimsical, "probable" status which the (Pythagorean)
Timaeus allows them.
ANDREW BARKER
University of Cambridge

22. The prime task of dialectic in the Republic is synthesis, bringing the results of the subsidiary
sciences together into a single system, ultimately through the understanding of ro ac'yaO6v. That an
understanding of the nature of unity itself, ro f', is crucial to this unifying project is nowhere
explicitly said (524E-525A has a merely arithmetical connotation for ro fV); though that knowledge
of a form is to be construed as knowledge of it as one is plain from, e.g., 524B-C. But in view of the
central importance given to mathematical studies here, and to the investigation of unity in later
dialogues (especially Parmenides, Sophist, Philebus), we might allow ourselves some small chink in
our orthodox skepticism concerning the well-known report of Aristoxenus, following Aristotle, that
Plato in some manner identified ro a&ya0ov and To fV (El. harm. 30).

This content downloaded from 99.238.17.184 on Sun, 15 Mar 2020 05:15:30 UTC
All use subject to https://about.jstor.org/terms